By Ross Douthat
This week my colleague David Brooks and I offered dispatches from two different futures: One in which Kamala Harris edges out Donald Trump for the presidency and one in which Trump is victorious. I wrote the “How Harris Wins” narrative, exploring a scenario in which the Democratic nominee succeeds in her effort to Marie Kondo-fy progressive politics, tidying things up by reducing the Democratic agenda to just a few popular components and letting that simplified, joy-sparking platform expose the internal tensions of the Republican Party’s coalition of the discontented.
That’s a vision of what could happen, and I think that Harris has a good chance to win in exactly the way that I describe. But if you forced me to place a bet on what will happen, my current expectations are closer to the scenario offered by my colleague — in which Trump, not Harris, is the next president of the United States.
One might argue that the safest way to bet is simply not to make one. As of this writing, Harris leads slightly in one of the popular betting markets, PredictIt, and Trump in another, Polymarket; in other words, for people making real wagers, it’s a toss-up. The RealClearPolitics polling average in Pennsylvania, the most likely decisive state, is a tie. Election forecaster Nate Silver’s complex model gives Trump a 60% chance of victory — but the forecasting at his former home, FiveThirtyEight, thinks Harris has a 57% chance of winning.
All this looks like the very definition of a coin-flip election. So why do I expect the coin to fall Trump’s way? Three reasons, none of them completely rigorous and all of them shadowed by the fact that I was wrong in 2016 (when I expected Trump to lose) and wrong in 2020 (when I expected Joe Biden to win more easily than he did), so I could simply be overcompensating for underestimating Trump’s chances in the past.
First, I think if Harris were on track to win, she would be leading more decisively at the moment. She has enjoyed an extended period of extraordinarily positive media coverage while the rival ticket flailed around trying to figure out an effective line of attack. She recently had the benefit of her party’s convention, which wrapped up Aug. 22 and was — in the press, at least — extremely positively received. And yet after those two boosts she still isn’t clearly ahead of Trump in the Electoral College race — which suggests that she probably now has more room to fall than rise.
Not that she will necessarily fall: It may be possible for her to sustain the media halo and the joyfully policy-light style for two more months, and in my essay on her path to victory that’s the future I assumed. But if the current dead heat is her ceiling, at least absent some dramatic change in the race, that’s enough reason for me to regard Trump as a very narrow favorite.
That is, as I understand it, part of why Silver’s projection now gives Trump a meaningful edge. What Silver’s calculations don’t include is an expectation of polling errors like the ones we saw, especially in some state polls, in 2016 and 2020, which led to Trump’s overperforming projections — and my second reason for betting on Trump is that I suspect he’ll slightly overperform again.
Silver explains why he isn’t forecasting such an error: Because the partisan valence of polling error varies from election to election, because pollsters have had four years to correct the problems that bedeviled them in 2020 and because two Trump elections is way too small a sample to assume that Trump will benefit from any error again.
All fair, but unscientifically I still suspect that pitting Trump’s unique coalition of the disaffected supporters against a Democratic coalition filled with institutionalist liberals who are overeager (especially now that Biden isn’t on the ticket) to tell pollsters how they’re voting creates survey problems that are hard for even the most careful and self-aware pollsters to fully overcome. Add in the murmurs from professionals and tea-leaf reading suggesting that the campaigns themselves don’t fully believe the numbers in the public polls, and I’m inclined to mentally add a point or two to Trump’s total in the averages — which again, would push him toward favorite territory.
Finally, like any analyst, I’m attached to my own theories, and my theory of this election before the great shake-up was that voters were alienated from Biden because he was seen as too liberal and not just because he was too old and that voter nostalgia for the Trump era had strengthened Trump’s position relative to four or eight years ago.
In that environment I was very doubtful that swapping Biden out for Harris, a figure burdened by his unpopular record and her own more liberal profile, would be enough to get the Democrats back to the (extremely narrow) edge they enjoyed in the key Electoral College states in 2020. And I still basically think that it won’t be enough — that notwithstanding Harris’ success in restoring Democratic enthusiasm, and notwithstanding her success so far in floating somewhere above and apart from her progressive record, she’s not a strong enough candidate to overcome the forces that gave Trump a lead in the first place.
So that’s why I’m still inclined to expect a Trump victory — for now, pending further developments or debate stage drama and with the awareness that this entire era is still designed to make fools of all political prognosticators.
Comments